Generalised Witt vectors (Q1261130)

A generalization of the ring of Witt vectors is the Burnside-Witt ring \(W_ G(A)\) whose elements are sequences of elements in \(A\) indexed by conjugacy classes of open subgroups of a profinite group \(G\), introduced by \textit{A. W. M. Dress} and \textit{C. Siebeneicher} [Adv. Math. 70, No. 1, 87-132 (1988; Zbl 0691.13026)]. This paper characterizes this ring algorithmically generalizing the carryless Witt vectors of \textit{N. Metropolis} and \textit{G.-G. Rota} [Adv. Math. 50, 95-125 (1983; Zbl 0545.05009)]. Here is an outline of the construction: For a \(G\)-set \(X\) (a group \(G\) acting on \(X\)) say that the action is essentially finite iff for any subgroup \(U\) of \(G\) of finite index the fixed set \(X^ U\) is finite; call it almost finite if also every element of \(X\) lies in a finite orbit. Define an equivalence relation on \(G\)-sets by \(X\sim Y\) iff for every \(U\) as above \(X^ U= Y^ U\). Disjoint union and Cartesian product endows the equivalences classes with a structure of a semiring satisfying the cancellation law, hence one can consider the associated ring, this is the Burnside ring \(\Omega(G)\). For a commutative ring \(A\) define a prestring on \(A\) to be an almost finite \(G\)-set and a map from orbits of \(X\) to \(A\). Define a string to be an equivalence class of prestrings for a natural equivalence relation. This way we obtain the semiring (with cancellation property) of strings and the associated ring is denoted by \(F_ G(A)\). Theorem 1 will show that \(F\) is the unique covariant functor such that \(F_ G(A)\) is the Burnside ring of \(G\)-strings with coefficients in \(A\). Defining now a morphism \(\psi_ G\) related to the ghost map (\(\varphi\) in the Dress-Siebeneicher paper) theorem 3 shows that \(W_ G(A)\) is isomorphic to \(F_ G(A)/\ker(\psi_ G)\) and proves the functoriality and some basic properties of this construction. A clearing algorithm analogous to that of Metropolis and Rota is used to get a set of generators for \(\ker(\psi_ G)\) and also to find coset representatives modulo this ideal. This simplifies the proof of the existence theorem. Here the group \(G\) has not to be profinite, however the construction depends only on the profinite completion of \(G\). For particular groups \(G\) one may get the classical Witt vectors, say for \(\mathbb{Z}/p\mathbb{Z}\) one obtains the classical truncated Witt vectors. It would be nice to have some applications of this construction, as one has for the original Witt vectors.